Quick Start¶
Create a Tree-based Tensor¶
You can create a tree-based tensor or a native tensor like the following example code.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 | import builtins import os from functools import partial import treetensor.torch as torch print = partial(builtins.print, sep=os.linesep) if __name__ == '__main__': t1 = torch.tensor([[1, 2, 3], [4, 5, 6]]) print('new native tensor:', t1) t2 = torch.tensor({ 'a': [1, 2, 3], 'b': {'x': [[4, 5], [6, 7]]}, }) print('new tree tensor:', t2) t3 = torch.randn(2, 3) print('new random native tensor:', t3) t4 = torch.randn({ 'a': (2, 3), 'b': {'x': (3, 4)}, }) print('new random tree tensor:', t4) |
The output should be like below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 | new native tensor: tensor([[1, 2, 3], [4, 5, 6]]) new tree tensor: <Tensor 0x7f3fb2f7d250> ├── 'a' --> tensor([1, 2, 3]) └── 'b' --> <Tensor 0x7f3fb2ed6a30> └── 'x' --> tensor([[4, 5], [6, 7]]) new random native tensor: tensor([[-1.2680, -0.2966, -0.4790], [ 1.8283, -0.6857, -0.9810]]) new random tree tensor: <Tensor 0x7f3fb2b69ee0> ├── 'a' --> tensor([[-1.4783, 2.4133, -1.8285], │ [-0.4211, -0.7495, -0.9413]]) └── 'b' --> <Tensor 0x7f3fb2f6b460> └── 'x' --> tensor([[ 0.3962, -2.1500, -0.3949, 1.1352], [-0.2198, -0.6111, 0.6990, -0.2641], [-0.5021, -0.3551, 0.6743, 1.1183]]) |